Freedom251 › Statistical Analysis Software
forestci
What forestci is
forestci is a free MIT-licensed Python package for adding variance calculations and confidence intervals to scikit-learn random-forest regression and classification objects. Its documented random_forest_error function calculates error bars from RandomForest estimators, including RandomForestRegressor and RandomForestClassifier. An optional calibrate flag applies Empirical Bayes calibration, although the documentation notes that calibration can reduce agreement with an empirical reference and is not always recommended. Proper confidence intervals require a large number of trees; the project recommends increasing the tree count toward convergence when calibration creates numerical errors. Version 0.9 supports scikit-learn versions 1.0 through 1.9. Installation uses `python -m pip install forestci`, with numpy, scipy, and scikit-learn listed as prerequisites. Examples cover regression with Auto MPG data and classification with synthetic data. The package runs through an API on Linux, welcomes pull requests, and directs issue reports to its issue log.
Who it suits
It suits Python developers using scikit-learn random forests who need error bars or confidence intervals for regression or classification. Users should be comfortable managing the listed Python dependencies and tree counts.
What it does well
- Free MIT-licensed package
- Works with regression and classification forests
- Optional Empirical Bayes calibration
- Supports scikit-learn 1.0 through 1.9
- Includes regression and classification examples
What to know first
- Requires numpy, scipy, and scikit-learn
- Confidence intervals need many trees
- Calibration may worsen empirical agreement
- Linux API package
Plans
| Plan | Billed and limits | Price |
|---|---|---|
| forestci | Open-source Python package · MIT license | Free |
As contrib.scikit-learn.org lists them, read 12 Oct 2026. The maker's own page has the final word.
Verdict
forestci adds uncertainty calculations directly to supported scikit-learn random-forest estimators. Its calibration option and tree-count requirements require judgment when interpreting results.
Facts from forestci's own pages
- API
- The documented `random_forest_error` function calculates error bars from scikit-learn RandomForest estimators. · contrib.scikit-learn.org, 12 Oct 2026
- Bayesian analysis
- No · contrib.scikit-learn.org, 23 Sept 2026
- Calibration
- The optional calibration routine attempts to extrapolate results for an infinite number of trees and may be unstable or cause numerical errors. · github.com, 12 Oct 2026
- Calibration limitation
- The documentation says calibration may worsen agreement with an empirical reference in some cases and is not universally recommended. · contrib.scikit-learn.org, 11 Oct 2026
- Compatibility
- The repository states that version 0.9 supports scikit-learn 1.0 through 1.9. · github.com, 12 Oct 2026
- Contribution
- The project welcomes contributions through pull requests and recommends discussing large changes in an issue first. · contrib.scikit-learn.org, 12 Oct 2026
- Contributions
- The documentation welcomes suggestions and contributions through pull requests and recommends discussing large changes in an issue first. · contrib.scikit-learn.org, 11 Oct 2026
- Core function
- The random_forest_error function calculates error bars from scikit-learn RandomForest estimators. · contrib.scikit-learn.org, 11 Oct 2026
- Dependencies
- The repository says users need NumPy, SciPy, and scikit-learn before installing forestci. · github.com, 12 Oct 2026
- Examples
- Examples demonstrate error bars for random forest regressors and classifiers using standard machine learning datasets. · contrib.scikit-learn.org, 12 Oct 2026
- Free plan
- Yes · contrib.scikit-learn.org, 23 Sept 2026
- Install
- The installation guide gives `python -m pip install forestci` as the install command. · contrib.scikit-learn.org, 12 Oct 2026
- Installation
- The installation guide gives `python -m pip install forestci` as the package installation command. · contrib.scikit-learn.org, 11 Oct 2026
- Integration
- Version 0.9 supports scikit-learn versions 1.0 through 1.9. · github.com, 11 Oct 2026
- License
- The repository identifies the software as licensed under the MIT open-source license. · github.com, 12 Oct 2026
- Method
- The package implements an algorithm developed by Wager and colleagues. · contrib.scikit-learn.org, 12 Oct 2026
- Purpose
- forestci adds variance calculations and confidence intervals to scikit-learn random forest regression and classification objects. · contrib.scikit-learn.org, 12 Oct 2026
- Regression models
- Yes · contrib.scikit-learn.org, 23 Sept 2026
- Scripting support
- Yes · contrib.scikit-learn.org, 23 Sept 2026
- Statistical tests
- No · contrib.scikit-learn.org, 23 Sept 2026
- Support
- The repository directs software issue reports to its issue log and lists email contacts for questions, suggestions, or feedback. · github.com, 12 Oct 2026
- Supported models
- The documentation describes confidence intervals for RandomForestRegressor and RandomForestClassifier objects. · contrib.scikit-learn.org, 11 Oct 2026
- Tree count
- The documentation says a large number of trees is needed to obtain proper confidence intervals. · github.com, 12 Oct 2026
Where forestci stands on Freedom251
- Statistical Analysis Software#42 of 106